## Statistical PERT – Normal Edition, New Version 2.0

I’m excited to announce that early in 2017, I will be releasing a new 2.0 version of the original Statistical PERT Excel workbooks and templates that I began creating in 2015. Version 2.0 will introduce **one-point**, probabilistic estimation by letting estimators choose a heuristic to calculate both the minimum and maximum point-estimates in a three-point estimate.

A **heuristic** is well-defined on Wikipedia as, “any approach to problem solving, learning, or discovery that employs a practical method not guaranteed to be optimal or perfect, but sufficient for the immediate goals. Where finding an optimal solution is impossible or impractical, heuristic methods can be used to speed up the process of finding a satisfactory solution. Heuristics can be mental shortcuts that ease the cognitive load of making a decision.”

What I’ve learned in the two years that I and others have used Statistical PERT is that sometimes people just don’t want to go through the mental work of figuring out a 3-point estimate for an uncertain outcome. Sometimes, it’s easier to start with a single estimate value and have the rest of the bell-shape curve implicitly created for you. And although that may not be as precise and optimal as specifying all three points of a 3-point estimate, using a heuristic to create two of the three points in a 3-point estimate may be “good enough” for project planning purposes.

In the 2017 release of Statistical PERT, estimators may choose to employ a simple heuristic that takes a single point-estimate — the most likely outcome — and reduces it by a specified percentage to create the **minimum** point-estimate, and increases the most likely outcome by a specified percentage to create the **maximum** point-estimate. SPERT estimators can choose separate heuristics for the minimum and maximum calculations.

For example, if I specify a value of 100 for the most likely outcome (the unit of measure doesn’t matter), I can say that the minimum point-estimate is 25% less than the most likely outcome, and the derived minimum would be 75. Then I can say that the maximum point-estimate is 50% greater than the most likely outcome, and the derived maximum would be 150. The resulting three-point estimate used to model the uncertainty would be (75, 100, 150). I only had to set the minimum and maximum heuristics once for the entire worksheet, and all my subsequent one-point estimates will use heuristics to automatically create the minimum and maximum point-estimates. Obviously, the math formulas involved are exceedingly simple, so the innovation here is mostly just about ease-of-use in creating three-point estimates, rather than technical innovation.

Estimators will be able to override heuristically-derived minimum and maximum point-estimates by just entering an actual, numeric value in place of the formula used to create minimum and maximum point-estimates. All SPERT download files will still offer a separate worksheet to enter 3-point estimates without using heuristics, too. Heuristics, then, are in addition to what is available in a Statistical PERT worksheet, not a replacement of anything currently available.

All existing Statistical PERT Excel downloads will be re-labeled as “Statistical PERT – Normal Edition” to reflect that all the currently-available downloads use Excel’s built-in **normal** distribution functions, NORM.DIST and NORM.INV.

Which brings me to the other exciting change for 2017….

## Statistical PERT – Beta Edition, New Version 1.0

It’s been one year since I began a new quest to create a simple way to create probabilistic estimates using the more-flexible **beta **distribution, which handles skewed uncertainties more accurately than the normal distribution does. I wanted to use Excel’s two beta distributions, BETA.DIST and BETA.INV in a spreadsheet that was just as easy to use as the original SPERT Excel spreadsheets, which hid the complexity of creating a standard deviation that the normal distribution functions require.

Excel’s beta distribution functions use two shape parameters, *alpha* and *beta*, which determine the precise shape of the beta curve. Specifying *alpha* and *beta* shape parameters is not at all intuitive, however, so the goal was to create a technique that figures out a good approximation of *alpha *and *beta* based solely upon the estimator’s three-point estimate and the estimator’s subjective opinion about how likely the most likely outcome really is (just like the original Statistical PERT spreadsheets using the normal distribution!).

The details behind Statistical PERT – Beta Edition are too much to include in this blog post, but you can see a development build of the forthcoming version 1.0 that will be released in early 2017, along with the version 2.0 of Statistical PERT – Normal Edition.

Together, both editions of Statistical PERT will let anyone easily make estimates using Excel’s built-in, statistical functions. And all Statistical PERT versions and editions will continue to be freely licensed under the GNU GPL, so you can download, use, modify and share any SPERT file like you’ve always been able to do.

Download a development-only build of Statistical PERT – Beta Edition here.